diff --git a/context/adr/002-naming-a-cell.md b/context/adr/002-naming-a-cell.md
index f777136c..80626411 100644
--- a/context/adr/002-naming-a-cell.md
+++ b/context/adr/002-naming-a-cell.md
@@ -226,3 +226,9 @@ learns no coordinate space, and an invalid operation is unconstructible rather
 than rejected. What the extension adds is that the enum is now **four** answers
 to "what may be done here", so `None` finally means only what this ADR always
 said it meant — no projected cell at all.
+
+**A fifth kind (2026-09-22, improvise-exdg).** The records-mode coordinate cell
+was the last `None` that named a visible cell. It is now `CellId::Coordinate`.
+The variant carries no operand, because no client composes a coordinate rename.
+The previous paragraph's "four answers" is therefore five. `None` now means
+only a padded or blanked slot. Reasoning: `context/code/protocol.md` §3.
diff --git a/context/code/clients.md b/context/code/clients.md
index f7ef1f19..06e08949 100644
--- a/context/code/clients.md
+++ b/context/code/clients.md
@@ -1746,6 +1746,31 @@ watchdog must be DISARMED and not merely pre-empted**: it fires later and would
 overwrite the better message with "did not answer" (PR #13 review). The watchdog
 is for silence only, and `setShellFailure` clears the timer.
 
+**Success is claimed by the first frame, and only if nothing failed first**
+(improvise-i0v9). The worker transport used to write `in-page worker · <doc>`
+the moment it spawned the worker — 13ms after navigation, before anything had
+answered — so a worker that could not start showed a success-looking status
+over a blank page. It now writes `starting the worker · <doc>…`, and
+`createShell().arm(message, { answer })` holds the claim until `sawFrame`. The
+rule the claim obeys is **the watchdog's own state**: it is written only if the
+watchdog is still armed when the first frame lands. That is not a convenience.
+`worker.js` posts a `ReadOnly` (or `SeedMissing`) notice straight after
+`start()` and **before** the Snapshot that answers the handshake, so an
+unconditional "first frame ⇒ say we are up" would overwrite the precise reason
+the notice just wrote. `fail` already disarms the watchdog, so "armed" answers
+"has anything been said since arming" without a second flag to keep in step.
+`fail` disarms **directly** rather than through `sawFrame`, or it would
+briefly flash the success text on its way to the failure. The starting text
+deliberately does not contain `in-page worker`, so the older specs asserting
+that substring now need the first frame to pass. The socket path passes no
+`answer`: `open` is already a real fact, and its status was never premature.
+
+One consequence changes what the watchdog guards. In the read-only case a
+Snapshot always follows the notice, and `sawFrame` disarms the watchdog either
+way, so a notice that forgot to disarm would **not** be overwritten eight
+seconds later. It is overwritten **at once**, by the claim. The eight-second
+overwrite survives only where no frame follows: `BootFailed`.
+
 **Did this socket ever open?** The page has to record that itself, because
 nothing on the close event carries it: `readyState` is already CLOSED by the time
 a close listener runs, whichever happened, and a refused connection and a
diff --git a/context/code/command.md b/context/code/command.md
index 3dd16fd3..a943e8bf 100644
--- a/context/code/command.md
+++ b/context/code/command.md
@@ -793,12 +793,40 @@ is gone rather than joined by a second one. A guard spelled here as
 copy of `switch_to_view`'s own condition, free to disagree with it the day that
 condition grows a clause.
 
-Not fixed, and reachable only by construction: `ViewBack` and `ViewForward` call
-`switch_to_view` just as unguardedly, pushing the current view onto the opposite
-stack before the switch they cannot be sure of. They are safe **because
-`ViewState::forget_view` purges both stacks when a view is deleted**, so a popped
-frame always names a live view — a guard living one level away from the hazard,
-which is the shape that keeps returning here.
+`ViewBack` and `ViewForward` have the same shape now (improvise-8nd9). They used to
+push the current view onto the opposite stack **before** a switch they could not be
+sure of, and were safe only because `ViewState::forget_view` purges both stacks when
+a view is deleted — a guard living one level away from the hazard. improvise-lrrx
+proved the guard incomplete: `App::install_workbook` (`LoadModel`,
+`AcceptWizardModel`) cleared `cursors` and not the history, so `<` after a load
+walked the **outgoing** document's frames, landing on a view of the same name in the
+incoming one that this client had never visited, or on a refused switch that still
+left a phantom frame behind. Both halves are fixed, so neither relies on the other:
+`install_workbook` clears both stacks beside `cursors`, for the same reason, and the
+two steps go through `stepped_to`, which switches, observes whether the document
+**accepted** the view (`App::switch_to_view` now returns that), and only then
+restores the frame's mode and hands back the frame just left. A frame the document
+refuses is consumed and leaves nothing.
+
+The observation is *accepted*, not *moved*, and that is the one place the sibling
+shapes differ. `SwitchView` asks "did `active_view` change" because asking for the
+view you are on is not a move. A history frame is a **place**, meaning a view and a
+mode, so a live frame naming the view you are on still restores its mode. Such
+frames exist: `forget_view` collapses `[A, B, A]` to `[A, A]`. The client's
+`reduce_view::stepped` also steps onto them unconditionally. A "moved" test
+would therefore drop a frame the client keeps, and the host and the client
+would disagree. The first version of this fix made exactly that mistake.
+`colon_view_back_respects_whether_history_exists` caught it, because its
+fixture is such a frame. `reduce_view`'s `stepped` is unchanged, because the
+client holds no `Workbook` to observe against. That leaves the host and the
+client able to disagree wherever a client's stack holds a frame the document
+would refuse. An earlier version of this paragraph said such a frame could not
+exist, because `ForgetView` purges deleted views and a load arrives as a fresh
+`Snapshot`. **That is false.** `project.rs::fresh_view` starts from
+`ViewState::default()` and folds `SwitchView(active)`, so every connected or
+rebased client gets a phantom `{Default, Normal}` back-frame, and `Default`
+need not exist in the document. That is improvise-lxg0, and until it is fixed
+the divergence is live rather than theoretical.
 
 ### `ReturnToNormalUnlessMoved` — observation, not prediction (improvise-toc)
 
@@ -2846,6 +2874,43 @@ flag: tidy-up goes where an abort cannot reach it.
 stays in `TileSelect` mode so the user can adjust several tiles. `axis: None` cycles;
 `axis: Some(a)` sets.
 
+**The tile bar offers `_Index` and `_Dim`, and that is deliberate (improvise-jgzw).**
+Every tile call site passes `Model::category_names()`, virtual categories included,
+so tile cursor 0 is `_Index`. The ticket asked whether that was a leak. It is not,
+for three reasons, each checked rather than argued:
+
+- **It is a decision the user already made.** improvise-kos: "`_Dim` and `_Index`
+  should work as pivot axes outside Records/Drill view". A category that works as a
+  pivot axis is one the axis picker offers.
+- **The tile bar is the only way out of a records layout that was saved.**
+  improvise-jgzw's "check what it does" found that hand-assigning `_Index` → Row and
+  `_Dim` → Column in a real view (tile bar or `:set-axis`) is **written** to the
+  `.improv`, because a real view persists every axis, and it reopens as records mode.
+  (`VIRTUAL_VIEW_NAMES`' "never-persisted `_Index`/`_Dim` on axes" is true only of
+  `_Records`/`_Drill`.) In that document, `R` (`toggle-records-mode`) emits
+  `ViewBack` on an empty back stack, so it leaves the layout as it was and sets the
+  status to "Pivot mode". That is filed as improvise-gj5p. Moving the two axes back through
+  the tile bar is what gets out. In the browser, which declines
+  `toggle-records-mode`, the tile bar is also the only way in, and improvise-bd9o
+  already made that route safe instead of closing it
+  (`the_tile_bar_route_into_records_mode_does_not_strand_the_user`).
+- **What filtering would cost was measured.** With `View::tileable_cats` filtering
+  on `CategoryKind::is_index_or_dim`, that bd9o test and the TUI's
+  `renders_each_tile_with_its_axis_label` fail, as well as the two characterization
+  tests here.
+
+What the default cursor does cost, stated: one cycle key on tile 0 puts `_Index`
+on Row alone, and in a pivot view that empties the grid (0 rows on
+`web/demo.improv`). The status line says `_Index → Row`, and three more presses
+undo it. If that turns out to matter, the fix is the order the tile bar lists
+categories in, not taking entries out of it.
+
+**`_Measure` must stay tileable.** It is on an axis in every shipped document, so
+`regular_category_names()`, the advice for *prompts*, is the wrong filter here.
+`the_tile_bar_offers_every_category_by_name_measure_and_records_axes_included`
+names the whole offered list with `_Measure` on a column, and fails if either
+decision changes.
+
 ### `cmd/effect_cmds.rs`
 
 The parseable model-mutation commands are thin `Cmd` wrappers around effects,
diff --git a/context/code/protocol.md b/context/code/protocol.md
index 3c5cc69e..f4cfb754 100644
--- a/context/code/protocol.md
+++ b/context/code/protocol.md
@@ -614,14 +614,37 @@ broader than its reason — improvise-gcvb, which also records the two things
 that make it more than a one-liner: an inserted `+cat` column means the TUI
 does **not** advance, and `AddCategory` can abort.
 
+### `CellId::Coordinate` — a drawn coordinate is not "no cell" (improvise-exdg)
+
+A records-mode coordinate cell (the `Region` column holding `North`) used to
+project `None`. That was a **third** meaning of `None`, next to the padded slot
+and the blanked position: a visible cell that holds text and whose write
+belongs to the host. So absence was encoding a permission, the same conflation
+`Create` removed for the creating columns (improvise-poia). The pinning test,
+`a_coordinate_column_is_projected_with_no_identity`, asserted the bug and is
+gone.
+
+`Coordinate` carries **nothing**. Editing a coordinate renames it across every
+measure cell in the row (gotcha 1), and no client composes that operation: it
+needs `record_col`, and the browser declines every records-mode edit before
+reaching a cell. Giving the variant a `RecordHash` would hand a client a
+surrogate with nothing to spend it on. The operand belongs with the command
+that consumes it, whenever that command exists. Nothing a client can do has
+changed. `render.rs::editable` is now an exhaustive `match` rather than a
+`matches!`, so the next variant has to be classified. `typed_cell_command`,
+`record_at_cursor` and `record_cell_at_cursor` still answer `None` for it.
+So paste onto a coordinate is still a silent `Upstream` in the browser. The
+TUI has declined it out loud since PR #184. That divergence is
+improvise-kkcy, and a named kind is what lets the browser decline deliberately
+now.
+
 ### `CellDisplay::id` — `None` is not "aggregate"
 
 That is `Some(CellId::Aggregate(..))`, and the distinction is the whole point
 of the enum. `None` means the slot holds **no projected cell**: a `GridCache`
-position padded by `fit_to_labels`, blanked by a reframing, or a
-records-mode column that is not a cell to write into at all (a category column
-renames a coordinate across the whole row, gotcha 1, which is a different
-intent from `SetCell`). Conflating the two either hides a cell or offers a
+position padded by `fit_to_labels`, or one blanked by a reframing. Every
+records-mode column is some kind of cell (`Record`, `Derived`, `Create` or
+`Coordinate`). Conflating the two either hides a cell or offers a
 write the type system exists to make unconstructible.
 
 ### `CellDisplay::edit_text` — the difference corrupts data
@@ -637,8 +660,8 @@ and an editor's seed is exactly a feedback path.
 It is the projection of `App::display_value_at_cursor`, which is what
 `EditOrDrill` seeds the TUI's buffer from, so both front ends start from one
 string rather than each deriving one. `None` where there is nothing to edit:
-the same set of positions `id` is `None` for, plus cells whose value is
-absent. An error cell carries none, because its text is a diagnostic and not
+positions whose `id` is `None`, plus pivot cells with no key. A records-mode
+cell always carries one, even an empty one. An error cell carries none, because its text is a diagnostic and not
 a value.
 
 A browser found this: `tests/lockstep.rs` drove `i` through the host and the
@@ -785,9 +808,16 @@ one**. `AxisFrame::flat` leaves `cats` empty on purpose, so it is not a way to
 smuggle a multi-level axis in flattened.
 
 `FrameGroup::collapsed` is on the wire rather than left to the client because
-it is a fact about the *view*, which only the server holds. `FrameGroup::cat`
-travels so a client can echo it back in `ToggleGroup` without ever working out
-which category a label came from — which slash-joined labels made impossible.
+it is a fact about the *view*, which only the server holds. There is no
+`FrameGroup::cat` any more (improvise-7xyx). Its doc comment promised that a
+client echoes it back in `ToggleGroup`, but no client ever did. `resolve.rs`
+has no `toggle-group-under-cursor` arm (`z` is on the still-needs-the-host
+list), and neither renderer read the field. A promise about an interaction no
+code performs is the stale-sibling defect, so the field went with it. It is
+not a loss for `z`. The TUI's `ToggleGroupAtCursor` also places the cursor on
+the layout the toggle *will produce*, which a thin client cannot compute. So
+composing `z` in a browser needs its own design, and the operand should come
+back with that design rather than wait for it.
 
 ### `AxisFrame`: four accessors that are **not for a draw loop**
 
@@ -1294,13 +1324,13 @@ frozen behaviour reads as decided afterwards and nothing says it was only frozen
 for the duration of the move.
 
 **The ticket's own proposed fix does not work, and that is worth more than the
-fix.** h0iv suggested comparing the `(cat, name)` pair, `FrameGroup` carrying
-`cat` precisely so a client need not work out which category a label came from.
+fix.** h0iv suggested comparing the `(cat, name)` pair; `FrameGroup` carried
+`cat` then (improvise-7xyx has since removed it).
 But its own reachability case is *one* category: `Month` grouped into `Q1`,
 nested under `Region`, giving `North/Q1` and `South/Q1` adjacent — the two
 headers are `(Month, Q1)` and `(Month, Q1)`, **equal as values**, so the pair
-separates nothing. `the_two_headers_are_equal_as_values_so_comparing_cat_and_name_cannot_split_them`
-pins that, so the next reader does not re-propose it. What separates them is
+separates nothing. `the_two_headers_are_equal_as_values_so_comparing_them_cannot_split_them`
+pins that as whole-value equality, so the next reader does not re-propose it. What separates them is
 that they are two different `FrameEntry::Group` entries, and `groups()` already
 borrows each from `self.entries` — so the comparison is `std::ptr::eq` on the
 header, which is the literal statement of "the same header opened both".
diff --git a/context/test-evidence/clients.md b/context/test-evidence/clients.md
index 9333ba89..144be8a5 100644
--- a/context/test-evidence/clients.md
+++ b/context/test-evidence/clients.md
@@ -3054,6 +3054,23 @@ The third is the gotcha 48 property, and it needs `timesArmed() > 0` first or it
 denies an unreachable event: the precise reason must still be on screen after
 the 8s window, not replaced by the watchdog. `page.clock` walks past it.
 
+**Strengthened by improvise-i0v9's second pass.** Every assertion above read the
+*final* `#status`, so two failures could still pass it. A watchdog that fired
+first and was then replaced by a late notice passed. So did a success claim
+shown before the notice. The test now records every status with `logStatus` and
+asserts, **before** the page clock is walked, that the log ends in `read-only`
+and never contained `did not answer` or the exact `in-page worker · <doc>`. A
+reason that arrived after the watchdog would sit in that log behind "did not
+answer"; that is the "well before 8s" property, stated without a wall-clock
+bound. Measured failing against `gh/main` at `da2e5ab`, whose log was
+`["in-page worker · two-tab-….improv", "read-only — …"]`: the premature claim
+the ticket measured at 13ms is still there once 1wft's notice exists. The
+brief's rival, `onNotice` calling `setShell` instead of `fail` (the watchdog not
+disarmed), fails it at the first `toContainText("read-only")`. The Snapshot's
+claim replaced the reason at once, which is why the eight-second form of this
+rival is now exercised by the invalid-`?doc=` spec instead (see
+`context/code/clients.md`, "The watchdog").
+
 It runs on Chromium and Firefox — the two engines with OPFS — rather than
 Chromium alone, because the exclusivity semantics are the engine's and the two
 word the refusal completely differently.
@@ -4473,6 +4490,38 @@ registers to agree; a TUI-only blank filter (`Text("")` to `None`) fails it on
 `+`, which is why that filter was removed rather than kept. The evidence entries
 for the host-side tests are in `command.md`.
 
+### the worker is not called up until its first frame has arrived
+
+improvise-i0v9. The worker transport wrote `in-page worker · <doc>` when it
+spawned the worker, before anything had answered. A worker that then failed
+therefore showed a success-looking status over a blank page. The spec records
+every `#status` with `logStatus` and requires the whole log to be exactly
+`["starting the worker · <doc>…", "in-page worker · <doc>"]`. It pins the order
+as well as the final text, since a final read cannot see a claim made too early.
+Measured failing against `gh/main` at `da2e5ab`. The rival, restoring the
+immediate `setShell(\`in-page worker · ${DOC}\`)` in `workerTransport`, fails it
+on the log's first entry. The final read is `toHaveText`, not `toContainText`.
+The older specs' `toContainText("in-page worker")` could not tell the starting
+text from the claim if the one contained the other, and the starting text
+deliberately does not.
+
+### only the first frame claims the peer answered, and not before it
+
+improvise-i0v9, the unit half of the spec above. `arm(message, { answer })`
+must write nothing until `sawFrame`, then write `answer` once, however many
+frames follow. Fails against `gh/main`'s `createShell`, which has no `answer`
+and so never writes the claim at all.
+
+### a reason given before the first frame is not replaced by the frame's claim
+
+The order `worker.js` really produces on a locked document: the `ReadOnly`
+notice is posted after `start()` and **before** the Snapshot. So `fail` then
+`sawFrame` must leave the notice's text as the only thing ever shown. The rival
+is the obvious fix, `sawFrame` writing `answer` unconditionally. It fails both
+this test and the one above: the first on its exact log, which is then written
+twice. This test passes against `gh/main` by construction, since no claim exists
+there. It is not evidence of the bug; it guards against the naive fix.
+
 ## Numbers across the serialized wire (improvise-6oy4)
 
 No manifest enabled serde_json's `float_roundtrip`, so its default f64 parser
diff --git a/context/test-evidence/command.md b/context/test-evidence/command.md
index d86e0500..f085ebb4 100644
--- a/context/test-evidence/command.md
+++ b/context/test-evidence/command.md
@@ -1978,12 +1978,27 @@ Two measure columns of one row are two cells of one record, and the next row is
 different record — the same shape the pivot case has, checked here because records
 mode reaches it by a different path.
 
-### a_coordinate_column_is_projected_with_no_identity
+### a_coordinate_cell_is_named_rather_than_left_as_no_cell
 
-A coordinate column is not a cell to write into: editing it renames that coordinate
-across *every* measure cell in the row (gotcha 1), which is a different intent from
-`SetCell`. So it is projected with no id at all, and a client cannot compose a
-write against it.
+improvise-exdg. This entry replaces `a_coordinate_column_is_projected_with_no_identity`,
+a test that **pinned the bug**. It asserted that a records-mode coordinate cell
+projects `id: None`. ADR 002 and gotcha 42 define `None` as "no projected cell",
+yet the Region column is drawn and holds text. So a client could not tell it
+from a padded slot. Absence was being used to encode a permission ("the host's"),
+which is the conflation ADR 002's rejected `Option<CellId>` draft was about.
+`improvise-poia` removed the same conflation for the creating columns.
+
+The cell now projects `CellId::Coordinate`. That keeps the old test's real
+property: it is still not a `SetCell` target, because `typed_cell_command`,
+`record_cell_at_cursor` and `render.rs::editable` all match `Record` alone.
+The test seats the cursor on the **second** row, so the first-row and
+first-item cases cannot coincide. It then enumerates every cell of the records
+grid and requires each to carry an id, which is the ticket's claim with no
+case picked by hand. Before the fix it failed at `a drawn coordinate projected
+as no cell`. Rival injected afterwards: `RecordCol::Category(_) => return None`
+restored. This test failed at the `Some(CellId::Coordinate)` assertion, and
+`a_creating_column_is_named_rather_than_left_as_no_cell` (whose `None`-count
+now expects zero) failed too.
 
 ### a_write_moves_the_version_under_the_clients_own_surrogate
 
@@ -3298,11 +3313,12 @@ aggregates, arriving a third time.
 
 The test walks **every** column of a records grid and counts, rather than
 asserting the two creating columns alone. That matters in the direction a
-weaker test fails: a producer that answered `Create` for everything, or that
-made the `Region` coordinate column nameable, would pass a two-position check
-and fails this one. The fault injected — returning `None` for both creating
-columns, which is the pre-fix code — reports three nameless columns where one
-is allowed.
+weaker test fails: a producer that answered `Create` for everything would pass
+a two-position check and fails this one. The fault injected — returning `None`
+for both creating columns, which is the pre-fix code — reported three nameless
+columns where one was allowed. Since improvise-exdg the coordinate column is
+named too (`CellId::Coordinate`), so the count of nameless columns is now zero.
+See `a_coordinate_cell_is_named_rather_than_left_as_no_cell`.
 
 ### a_records_coordinate_is_text_however_it_is_spelled
 
@@ -5134,6 +5150,57 @@ the list. Its seed is exact (`"007"`), but `typed()` classifies the committed
 text as `Number(7.0)`. Probed here: both tests fail on it in both modes. That
 is the classifier, not the seed, and it is improvise-ixxx.
 
+### loading_a_document_forgets_the_outgoing_documents_view_history
+
+improvise-lrrx: `App::install_workbook` cleared `cursors` but not
+`view_back_stack`/`view_forward_stack`, so `<` after `LoadModel` walked the
+outgoing document's views inside the incoming one. The history is two deep
+(`Default`, `X`, then on `Z`), and the incoming document **has** a view named `X`
+but opens on `Y`. Without that, a stale frame is a refused switch and looks the
+same as "nothing happened" (the ticket's own fixture warning). Against the unfixed
+code, `<` landed on `X`, a view this client never visited in the new document.
+Rival injected: the two `clear()` lines removed from `install_workbook`, with the
+improvise-8nd9 fix still in place. This test and its wizard twin failed at the
+`view-back walked the outgoing document's history` assertion.
+
+### accepting_an_import_forgets_the_outgoing_documents_view_history
+
+improvise-lrrx, through the second door into `install_workbook`. A
+wizard-built workbook has only `Default`, so the test creates an `X` in it after
+accepting. Without that, the stale `X` frame would be refused and could not be
+told apart from a cleared history. Same rival and same failure as the load test.
+
+### a_step_back_the_document_refuses_leaves_no_frame_behind
+
+improvise-8nd9: `ViewBack` pushed the current view onto the forward stack before
+the guarded `switch_to_view`, so a frame naming a view the document lacks left a
+phantom behind, and `,` would then "go forward" to where it already was. The
+dead frame is pushed by hand, standing for any path that forgets
+`forget_view`, with a live frame under it and the frame's mode set to `Help` so
+that a restored mode is observable. Two rivals. First, the original code, which
+pushes first and restores the mode unconditionally: it fails at `nothing moved,
+so no mode`. Second, a push-first version that restores the mode only on a move,
+so the mode assertion cannot hide the stack assertion: it fails at `a refused
+step left X on the forward stack`. The last two assertions check that the live
+frame beneath still works, so a fix that drops the whole stack cannot pass.
+
+### a_step_back_to_the_view_you_are_on_is_the_step_the_client_folds
+
+improvise-8nd9. This test holds the fix to the client's `reduce_view::stepped`.
+A live frame naming the view you are on (a `forget_view` collapse leaves one)
+is still a step. It restores its mode and leaves a forward frame, exactly as
+the client folds it. Rival injected: `stepped_to` observing whether
+`active_view` *changed* rather than whether the document *accepted* the view.
+That was the first version of this fix. This test failed at the
+mode assertion, as did
+`command::cmd::text_buffer::tests::colon_view_back_respects_whether_history_exists`.
+That test's fixture is such a frame, and it is what caught the first version.
+
+### a_step_forward_the_document_refuses_leaves_no_frame_behind
+
+The `ViewForward` half of the test above, with the stacks' roles swapped. It fails
+against the unfixed code at `nothing moved, so no mode`.
+
 ### yy_then_p_in_records_mode_copies_between_real_records
 
 improvise-6p8q. At `gh/main` `9e78051` it failed at the yank: `left: None,
@@ -5246,3 +5313,27 @@ refusal is recorded and the mode is not `Quit`. `import-json` failed first,
 before `refuse_and_stop` was applied to its arms. For `load`, the injected fault
 is removing its abort (`refuse_and_stop` back to `app.refuse`), and the test
 then fails with "`load absent.improv . q` quit having loaded nothing".
+
+## `command/cmd/tile.rs` — the tile bar's offered set, by name (improvise-jgzw)
+
+### the_tile_bar_offers_every_category_by_name_measure_and_records_axes_included
+
+This test pins a decision, not a bug. The argument is in `context/code/command.md`
+under `cmd/tile.rs`. Before this test, no tile assertion named the whole offered
+set. The two older ones named only the category at cursor 0, so a filter that
+changed what the bar offers past the cursor was invisible. The fixture puts
+`_Measure` on **Column**, as every shipped document does. The test then walks the
+cursor past the end and asserts the exact list
+`[_Index, _Dim, _Measure, Type, Month]`.
+
+Rivals, injected, full workspace run:
+
+- **`regular_category_names()` in `TileAxisOp`**, the obvious "no virtual
+  categories in pickers" fix, which drops `_Measure` too. It fails this test
+  (`["Type", "Month"]`), the two characterization tests above it, and
+  `lockstep.rs::a_locally_resolved_tile_axis_change_agrees_with_the_host`.
+- **`View::tileable_cats` filtering `is_index_or_dim`**, the ticket's option (b).
+  With the list asserted this test fails by construction, so the point is what
+  *else* fails: `renders_each_tile_with_its_axis_label` (TUI) and
+  `the_tile_bar_route_into_records_mode_does_not_strand_the_user`. The second one
+  is improvise-bd9o's own test of the route this decision keeps.
diff --git a/context/test-evidence/protocol.md b/context/test-evidence/protocol.md
index a0e901ef..e1dc17a6 100644
--- a/context/test-evidence/protocol.md
+++ b/context/test-evidence/protocol.md
@@ -557,7 +557,7 @@ the evidence that nothing was relying on the permissive parse.
 
 ### render::tests::two_runs_under_two_headers_of_one_name_are_two_runs
 
-### render::tests::the_two_headers_are_equal_as_values_so_comparing_cat_and_name_cannot_split_them
+### render::tests::the_two_headers_are_equal_as_values_so_comparing_them_cannot_split_them
 
 improvise-h0iv. `group_runs` coalesced adjacent positions by group **name**, so
 two runs under two different headers sharing a name read as one and both front
@@ -571,10 +571,12 @@ and nothing ungrouped between the runs. Before the fix it returns
 at the precondition rather than pass with a single run.
 
 The second test exists because the ticket proposed the wrong fix and a reader
-will propose it again. h0iv says `FrameGroup` carries `cat` and "comparing the
-pair would fix it" — but its own reachability case is a single category nested
-under an outer one, so both headers are `(Month, Q1)`, equal as values. The test
-asserts that equality directly. What actually separates them is that they are
+will propose it again. h0iv said `FrameGroup` carried `cat` and that "comparing
+the pair would fix it". But its own reachability case is a single category
+nested under an outer one, so both headers are `(Month, Q1)`, equal as values.
+The test asserts that equality directly. Since improvise-7xyx removed
+`FrameGroup::cat`, it asserts equality of the whole header, which is stronger:
+no field-wise comparison can separate two equal values. What actually separates them is that they are
 two distinct entries in `self.entries`, which `std::ptr::eq` on the borrowed
 header states exactly.
 
diff --git a/crates/improvise-command/src/command/cmd/tile.rs b/crates/improvise-command/src/command/cmd/tile.rs
index 1ff05418..c4a71cf9 100644
--- a/crates/improvise-command/src/command/cmd/tile.rs
+++ b/crates/improvise-command/src/command/cmd/tile.rs
@@ -59,6 +59,26 @@ mod tests {
         assert!(effects.is_empty());
     }
 
+    /// improvise-jgzw: `_Index`/`_Dim` stay tileable on purpose, and `_Measure` on
+    /// a column must too — `context/test-evidence/command.md`.
+    #[test]
+    fn the_tile_bar_offers_every_category_by_name_measure_and_records_axes_included() {
+        let mut m = two_cat_model();
+        m.active_view_mut().set_axis("_Measure", Axis::Column);
+        let layout = make_layout(&m);
+        let reg = make_registry();
+        let mut ctx = make_ctx(&m, &layout, &reg);
+        let mut offered = Vec::new();
+        for idx in 0..8 {
+            ctx.tile_cat_idx = idx;
+            let dbg = effects_debug(&TileAxisOp { axis: None }.execute(&ctx));
+            if let Some(rest) = dbg.split("CycleAxis(Active, \"").nth(1) {
+                offered.push(rest.split('"').next().unwrap_or_default().to_string());
+            }
+        }
+        assert_eq!(offered, ["_Index", "_Dim", "_Measure", "Type", "Month"]);
+    }
+
     #[test]
     fn move_tile_cursor_right() {
         let m = two_cat_model();
diff --git a/crates/improvise-command/src/ui/app.rs b/crates/improvise-command/src/ui/app.rs
index 565280a6..4969f2cb 100644
--- a/crates/improvise-command/src/ui/app.rs
+++ b/crates/improvise-command/src/ui/app.rs
@@ -444,6 +444,9 @@ impl App {
     pub fn install_workbook(&mut self, workbook: Workbook) {
         self.view_state.active_view = workbook.active_view.clone();
         self.view_state.cursors.clear();
+        // The history names the outgoing document's views too (improvise-lrrx).
+        self.view_state.view_back_stack.clear();
+        self.view_state.view_forward_stack.clear();
         self.model_state.workbook = workbook;
         // Every client's mirror is scoped to the outgoing document too, and an
         // effect cannot reach the session list. The generation is how it hears.
@@ -581,12 +584,14 @@ impl App {
             .flat_map(move |r| (0..layout.col_count()).filter_map(move |c| layout.cell_key(r, c)))
     }
 
-    /// Point **this client's** slice at `name`, if the document has such a view.
-    /// The document's own default stays put (improvise-jytm, gotcha 29).
-    pub fn switch_to_view(&mut self, name: &str) {
-        if self.model_state.workbook.views.contains_key(name) {
+    /// Point **this client's** slice at `name`, if the document has such a view,
+    /// and say whether it did. The document's default stays put (gotcha 29).
+    pub fn switch_to_view(&mut self, name: &str) -> bool {
+        let accepted = self.model_state.workbook.views.contains_key(name);
+        if accepted {
             self.view_state.active_view = name.to_string();
         }
+        accepted
     }
 
     /// What `name` names in this document, falling back to the default for a name
diff --git a/crates/improvise-command/src/ui/effect.rs b/crates/improvise-command/src/ui/effect.rs
index 9bc78b34..18528e69 100644
--- a/crates/improvise-command/src/ui/effect.rs
+++ b/crates/improvise-command/src/ui/effect.rs
@@ -388,19 +388,30 @@ impl Effect for SwitchView {
     }
 }
 
+/// Switch to `frame`'s view and restore its mode, returning the frame just left
+/// — or `None`, touching nothing, if the document refused the view.
+fn stepped_to(app: &mut App, frame: ViewFrame) -> Option<ViewFrame> {
+    let left = ViewFrame {
+        view_name: app.view_state.active_view.clone(),
+        mode: app.view_state.mode.clone(),
+    };
+    app.switch_to_view(&frame.view_name).then(|| {
+        app.view_state.mode = frame.mode;
+        left
+    })
+}
+
 /// Go back in view history (pop back stack, push current to forward stack).
 #[derive(Debug)]
 pub struct ViewBack;
 impl Effect for ViewBack {
     fn apply(&self, app: &mut App) {
-        if let Some(frame) = app.view_state.view_back_stack.pop() {
-            let current = app.view_state.active_view.clone();
-            app.view_state.view_forward_stack.push(ViewFrame {
-                view_name: current,
-                mode: app.view_state.mode.clone(),
-            });
-            app.switch_to_view(&frame.view_name);
-            app.view_state.mode = frame.mode;
+        // Observed, as `SwitchView` is: a frame the document refuses is dropped
+        // and leaves nothing behind (improvise-8nd9, `context/code/command.md`).
+        if let Some(frame) = app.view_state.view_back_stack.pop()
+            && let Some(left) = stepped_to(app, frame)
+        {
+            app.view_state.view_forward_stack.push(left);
         }
     }
     fn kind(&self) -> EffectKind {
@@ -413,14 +424,11 @@ impl Effect for ViewBack {
 pub struct ViewForward;
 impl Effect for ViewForward {
     fn apply(&self, app: &mut App) {
-        if let Some(frame) = app.view_state.view_forward_stack.pop() {
-            let current = app.view_state.active_view.clone();
-            app.view_state.view_back_stack.push(ViewFrame {
-                view_name: current,
-                mode: app.view_state.mode.clone(),
-            });
-            app.switch_to_view(&frame.view_name);
-            app.view_state.mode = frame.mode;
+        // The mirror of `ViewBack`, observed the same way (improvise-8nd9).
+        if let Some(frame) = app.view_state.view_forward_stack.pop()
+            && let Some(left) = stepped_to(app, frame)
+        {
+            app.view_state.view_back_stack.push(left);
         }
     }
     fn kind(&self) -> EffectKind {
@@ -1925,6 +1933,169 @@ mod tests {
         assert_eq!(app.view_state.active_view, before);
     }
 
+    /// A history two deep through `X` and `Z`: back is `[Default, X]`.
+    fn app_with_history(storage: Box<dyn crate::storage::Storage>) -> App {
+        let mut wb = Workbook::new("Outgoing");
+        wb.create_view("X");
+        wb.create_view("Z");
+        let mut app = App::with_storage(wb, None, storage);
+        for name in ["X", "Z"] {
+            SwitchView(name.to_string()).apply(&mut app);
+        }
+        assert_eq!(app.view_state.view_back_stack.len(), 2, "setup: two deep");
+        app
+    }
+
+    /// Asserts `<` and `,` leave `view` put and the history empty.
+    fn assert_history_went_with_the_document(app: &mut App, view: &str) {
+        assert_eq!(app.view_state.active_view, view, "setup: opened on {view}");
+        for (name, step) in [
+            ("view-back", Box::new(ViewBack) as Box<dyn Effect>),
+            ("view-forward", Box::new(ViewForward)),
+        ] {
+            step.apply(app);
+            assert_eq!(
+                app.view_state.active_view, view,
+                "{name} walked the outgoing document's history in the incoming one"
+            );
+        }
+        assert!(app.view_state.view_back_stack.is_empty());
+        assert!(app.view_state.view_forward_stack.is_empty());
+    }
+
+    /// improvise-lrrx. Evidence: `context/test-evidence/command.md`.
+    #[test]
+    fn loading_a_document_forgets_the_outgoing_documents_view_history() {
+        use crate::storage::{MemoryFs, Storage};
+        let fs = MemoryFs::new();
+        // The incoming document HAS a view named `X`, and opens somewhere else.
+        let mut incoming = Workbook::new("Incoming");
+        incoming.create_view("X");
+        incoming.create_view("Y");
+        incoming.active_view = "Y".to_string();
+        let text = crate::persistence::format_md(&incoming);
+        fs.write(std::path::Path::new("b.improv"), text.as_bytes())
+            .unwrap();
+        let mut app = app_with_history(Box::new(fs));
+
+        LoadModel("b.improv".into()).apply(&mut app);
+        assert_eq!(
+            app.model_state.workbook.model.name, "Incoming",
+            "setup: loaded"
+        );
+        assert_history_went_with_the_document(&mut app, "Y");
+    }
+
+    /// improvise-lrrx: the wizard is the other door into `install_workbook`.
+    #[test]
+    fn accepting_an_import_forgets_the_outgoing_documents_view_history() {
+        let mut app = app_with_history(Box::new(crate::storage::MemoryFs::new()));
+        let mut wizard = crate::import::wizard::ImportWizard::new(serde_json::json!([
+            {"region": "East", "revenue": 100.0},
+            {"region": "West", "revenue": 200.0},
+        ]));
+        wizard.advance();
+        wizard.advance();
+        wizard.build();
+        assert!(wizard.built_summary().is_some(), "setup: a built workbook");
+        app.wizard = Some(wizard);
+
+        AcceptWizardModel.apply(&mut app);
+        // The new document gains an `X` of its own, so a stale frame could land.
+        app.model_state.workbook.create_view("X");
+        let opened = app.view_state.active_view.clone();
+        assert_ne!(opened, "X", "setup: the import must not open on X");
+        assert_history_went_with_the_document(&mut app, &opened);
+    }
+
+    /// improvise-8nd9: a frame naming a view the document lacks is refused, and
+    /// a refused step leaves no frame behind. See `context/test-evidence/command.md`.
+    #[test]
+    fn a_step_back_the_document_refuses_leaves_no_frame_behind() {
+        let mut app = test_app();
+        app.model_state.workbook.create_view("X");
+        SwitchView("X".to_string()).apply(&mut app);
+        // A path that forgot `forget_view`: the history names a dead view.
+        app.view_state.view_back_stack.push(ViewFrame {
+            view_name: "Gone".to_string(),
+            mode: AppMode::Help,
+        });
+
+        ViewBack.apply(&mut app);
+        assert_eq!(
+            app.view_state.active_view, "X",
+            "setup: the step is refused"
+        );
+        assert_eq!(
+            app.view_state.mode,
+            AppMode::Normal,
+            "nothing moved, so no mode"
+        );
+        assert!(
+            app.view_state.view_forward_stack.is_empty(),
+            "a refused step left X on the forward stack: `,` would go nowhere"
+        );
+        ViewBack.apply(&mut app);
+        assert_eq!(
+            app.view_state.active_view, "Default",
+            "the live frame survives"
+        );
+        assert_eq!(app.view_state.view_forward_stack.len(), 1);
+    }
+
+    /// improvise-8nd9: a live frame naming the view you are on is still a step, as
+    /// it is in the client's `reduce_view`. See `context/test-evidence/command.md`.
+    #[test]
+    fn a_step_back_to_the_view_you_are_on_is_the_step_the_client_folds() {
+        let mut app = test_app();
+        app.model_state.workbook.create_view("X");
+        SwitchView("X".to_string()).apply(&mut app);
+        app.view_state.view_back_stack.push(ViewFrame {
+            view_name: "X".to_string(),
+            mode: AppMode::Help,
+        });
+        let mut mirror = app.view_state.clone();
+
+        ViewBack.apply(&mut app);
+        improvise_protocol::reduce_view(&mut mirror, &improvise_protocol::Command::ViewBack);
+        assert_eq!(
+            app.view_state.mode,
+            AppMode::Help,
+            "the frame's mode is restored"
+        );
+        assert_eq!(app.view_state.view_back_stack, mirror.view_back_stack);
+        assert_eq!(app.view_state.view_forward_stack, mirror.view_forward_stack);
+        assert_eq!(app.view_state.mode, mirror.mode);
+    }
+
+    /// improvise-8nd9: the forward half of the step above.
+    #[test]
+    fn a_step_forward_the_document_refuses_leaves_no_frame_behind() {
+        let mut app = test_app();
+        app.model_state.workbook.create_view("X");
+        SwitchView("X".to_string()).apply(&mut app);
+        ViewBack.apply(&mut app);
+        app.view_state.view_forward_stack.push(ViewFrame {
+            view_name: "Gone".to_string(),
+            mode: AppMode::Help,
+        });
+
+        ViewForward.apply(&mut app);
+        assert_eq!(app.view_state.active_view, "Default", "setup: refused");
+        assert_eq!(
+            app.view_state.mode,
+            AppMode::Normal,
+            "nothing moved, so no mode"
+        );
+        assert!(
+            app.view_state.view_back_stack.is_empty(),
+            "a refused step left Default on the back stack: `<` would go nowhere"
+        );
+        ViewForward.apply(&mut app);
+        assert_eq!(app.view_state.active_view, "X", "the live frame survives");
+        assert_eq!(app.view_state.view_back_stack.len(), 1);
+    }
+
     #[test]
     fn create_and_delete_view() {
         let mut app = test_app();
diff --git a/crates/improvise-command/src/ui/project.rs b/crates/improvise-command/src/ui/project.rs
index 2a433d94..ff23ca77 100644
--- a/crates/improvise-command/src/ui/project.rs
+++ b/crates/improvise-command/src/ui/project.rs
@@ -222,7 +222,6 @@ fn axis_frame(
                     pending.clear();
                 }
                 pending.push(FrameEntry::Group(FrameGroup {
-                    cat: cat_name.clone(),
                     name: group_name.clone(),
                     collapsed: view.is_group_collapsed(cat_name, group_name),
                 }));
@@ -324,7 +323,7 @@ fn cell_id(app: &App, layout: &GridLayout, row: usize, col: usize) -> Option<Cel
             RecordCol::NewCategory => return Some(CellId::Create(CreateKind::Category)),
             // Renaming a coordinate moves every measure cell in the row, which is a
             // different intent from `SetCell` and still the host's (gotcha 1).
-            RecordCol::Category(_) => return None,
+            RecordCol::Category(_) => return Some(CellId::Coordinate),
         };
         return Some(CellId::Record(RecordCell {
             record: RecordHash::of(record),
@@ -493,7 +492,10 @@ pub fn published_basis(app: &App, grid: &GridCache) -> Basis {
         .flatten()
         .filter_map(|cell| match cell.id.as_ref()? {
             CellId::Record(rc) => Some((rc.record, app.record_version_of(rc.record))),
-            CellId::Aggregate(_) | CellId::Derived { .. } | CellId::Create(_) => None,
+            CellId::Aggregate(_)
+            | CellId::Derived { .. }
+            | CellId::Create(_)
+            | CellId::Coordinate => None,
         })
         .collect();
     basis::published(app.document_generation(), app.layout_version(), records)
@@ -4853,20 +4855,6 @@ mod cell_identity {
         );
     }
 
-    /// A coordinate column is not a cell to write into (gotcha 1), so it is projected
-    /// with no id and a client cannot compose a write against it.
-    #[test]
-    fn a_coordinate_column_is_projected_with_no_identity() {
-        let app = records_app();
-        let region = column_named(&app, "Region");
-        assert_eq!(
-            grid_of(&app, WHOLE)
-                .cell(0, region)
-                .and_then(|c| c.id.clone()),
-            None
-        );
-    }
-
     /// A document where `Profit` is BOTH an explicit `_Measure` item and a
     /// formula target — which the `+` column lets a user build.
     fn shadowed_measure_records_app() -> App {
@@ -5138,10 +5126,9 @@ mod cell_identity {
         let ids: Vec<Option<CellId>> = (0..app.layout.col_count())
             .map(|c| grid.cell(0, c).and_then(|cell| cell.id.clone()))
             .collect();
-        assert_eq!(
-            ids.iter().filter(|id| id.is_none()).count(),
-            1,
-            "only the one Region coordinate column may be nameless: {ids:?}"
+        assert!(
+            ids.iter().all(Option::is_some),
+            "no records column is nameless since improvise-exdg: {ids:?}"
         );
         assert_eq!(
             ids.iter()
@@ -5162,6 +5149,30 @@ mod cell_identity {
         );
     }
 
+    /// improvise-exdg: a coordinate cell is drawn and holds text, so it is not
+    /// "no cell". `context/test-evidence/command.md`.
+    #[test]
+    fn a_coordinate_cell_is_named_rather_than_left_as_no_cell() {
+        let app = records_app();
+        let grid = grid_of(&app, WHOLE);
+        let region = column_named(&app, "Region");
+        let north = grid.cell(1, region).expect("the second row's coordinate");
+        assert_eq!(north.text, "North", "setup: a visible coordinate cell");
+        assert_eq!(
+            north.id,
+            Some(CellId::Coordinate),
+            "a drawn coordinate is named"
+        );
+        let nameless: Vec<(usize, usize)> = (0..app.layout.row_count())
+            .flat_map(|r| (0..app.layout.col_count()).map(move |c| (r, c)))
+            .filter(|&(r, c)| grid.cell(r, c).is_some_and(|cell| cell.id.is_none()))
+            .collect();
+        assert!(
+            nameless.is_empty(),
+            "every records cell is something: {nameless:?}"
+        );
+    }
+
     /// The version counter is keyed by the surrogate the client holds, so one
     /// identity answers ADR 002's question and ADR 001's.
     #[test]
diff --git a/crates/improvise-protocol/src/render.rs b/crates/improvise-protocol/src/render.rs
index 647e9bb1..6397136e 100644
--- a/crates/improvise-protocol/src/render.rs
+++ b/crates/improvise-protocol/src/render.rs
@@ -55,6 +55,9 @@ pub enum CellId {
     /// A records-mode creating column: typing a **name** into it adds that
     /// measure or category and writes no cell at all (gotcha 1, improvise-poia).
     Create(CreateKind),
+    /// A records-mode coordinate. Renaming it moves every measure cell in the
+    /// row (gotcha 1), which no client composes: `protocol.md` §3 (improvise-exdg).
+    Coordinate,
 }
 
 /// What a creating column creates. The name typed into it is the whole
@@ -194,9 +197,6 @@ pub const MAX_ROW_HEADER_W: u16 = 24;
 /// because it is a fact about the *view*, which only the server holds.
 #[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
 pub struct FrameGroup {
-    /// The category the group belongs to, echoed back in `ToggleGroup` so a
-    /// client never works out which category a label came from.
-    pub cat: String,
     pub name: String,
     pub collapsed: bool,
 }
@@ -1137,7 +1137,6 @@ mod tests {
                 FrameEntry::Item(vec!["A".into(), "x".into()]),
                 FrameEntry::Item(vec!["A".into(), "y".into()]),
                 FrameEntry::Group(FrameGroup {
-                    cat: "Type".into(),
                     name: "Late".into(),
                     collapsed: true,
                 }),
@@ -1150,7 +1149,6 @@ mod tests {
     /// ungrouped between them — improvise-h0iv's reachable shape.
     fn repeated_group_frame() -> AxisFrame {
         let q1 = || FrameGroup {
-            cat: "Month".into(),
             name: "Q1".into(),
             collapsed: false,
         };
@@ -1176,7 +1174,6 @@ mod tests {
             cats: vec!["Month".into()],
             entries: vec![
                 FrameEntry::Group(FrameGroup {
-                    cat: "Month".into(),
                     name: "Q1".into(),
                     collapsed: false,
                 }),
@@ -1214,7 +1211,6 @@ mod tests {
             cats: vec!["Month".into()],
             entries: vec![
                 FrameEntry::Group(FrameGroup {
-                    cat: "Month".into(),
                     name: "Q1".into(),
                     collapsed: true,
                 }),
@@ -1266,7 +1262,7 @@ mod tests {
     /// The pair `FrameGroup` carries cannot separate them — h0iv's own proposed
     /// fix. See `context/test-evidence/protocol.md`.
     #[test]
-    fn the_two_headers_are_equal_as_values_so_comparing_cat_and_name_cannot_split_them() {
+    fn the_two_headers_are_equal_as_values_so_comparing_them_cannot_split_them() {
         let frame = repeated_group_frame();
         let headers: Vec<&FrameGroup> = frame
             .entries
@@ -1277,10 +1273,9 @@ mod tests {
             })
             .collect();
         assert_eq!(
-            (&headers[0].cat, &headers[0].name),
-            (&headers[1].cat, &headers[1].name),
-            "both headers name the same category AND the same group, so the \
-             ticket's (cat, name) comparison would still merge them"
+            headers[0], headers[1],
+            "both headers are the same group in the same state, so no \
+             comparison of their fields can split them"
         );
     }
 
@@ -1344,7 +1339,7 @@ mod tests {
         assert_eq!(frame.item(2), Some(&vec!["B".to_string(), "z".to_string()]));
         assert!(frame.group_of(0).is_none(), "[A,x] is in no group");
         let group = frame.group_of(2).expect("[B,z] sits under Late");
-        assert_eq!((group.cat.as_str(), group.name.as_str()), ("Type", "Late"));
+        assert_eq!(group.name, "Late");
         assert!(group.collapsed, "the collapse state travels with the group");
     }
 
diff --git a/crates/improvise-protocol/src/resolve.rs b/crates/improvise-protocol/src/resolve.rs
index 3a86f346..0e5d0c6e 100644
--- a/crates/improvise-protocol/src/resolve.rs
+++ b/crates/improvise-protocol/src/resolve.rs
@@ -756,6 +756,7 @@ fn edit_or_drill(state: &ClientState, mode: AppMode) -> Option<Vec<Command>> {
         // Both creating columns are records-only, so the guard above takes
         // them; declining rather than `None` keeps that a door, not a hole.
         CellId::Create(_) => Some(vec![Command::SetStatus(not_yet("creating a column"))]),
+        CellId::Coordinate => Some(vec![Command::SetStatus(not_yet("renaming a coordinate"))]),
     }
 }
 
@@ -949,7 +950,9 @@ fn advanced(state: &ClientState, edit_mode: AppMode, advance: Advance) -> Option
 fn record_at_cursor(state: &ClientState) -> Option<crate::RecordHash> {
     match cell_at_cursor(state)?.id.as_ref()? {
         CellId::Record(cell) => Some(cell.record),
-        CellId::Aggregate(_) | CellId::Derived { .. } | CellId::Create(_) => None,
+        CellId::Aggregate(_) | CellId::Derived { .. } | CellId::Create(_) | CellId::Coordinate => {
+            None
+        }
     }
 }
 
@@ -1016,7 +1019,9 @@ fn model(state: &ClientState, cmd: ModelCommand) -> Option<Command> {
 fn record_cell_at_cursor(state: &ClientState) -> Option<RecordCell> {
     match cell_at_cursor(state)?.id.as_ref()? {
         CellId::Record(cell) => Some(cell.clone()),
-        CellId::Aggregate(_) | CellId::Derived { .. } | CellId::Create(_) => None,
+        CellId::Aggregate(_) | CellId::Derived { .. } | CellId::Create(_) | CellId::Coordinate => {
+            None
+        }
     }
 }
 
diff --git a/crates/improvise-wasm-client/src/lib.rs b/crates/improvise-wasm-client/src/lib.rs
index 6b696c1f..e02f7cf9 100644
--- a/crates/improvise-wasm-client/src/lib.rs
+++ b/crates/improvise-wasm-client/src/lib.rs
@@ -180,7 +180,10 @@ impl Client {
     pub fn record_cell_at(&self, row: usize, col: usize) -> Option<RecordCell> {
         match self.state.grid.cell(row, col)?.id.clone()? {
             CellId::Record(cell) => Some(cell),
-            CellId::Aggregate(_) | CellId::Derived { .. } | CellId::Create(_) => None,
+            CellId::Aggregate(_)
+            | CellId::Derived { .. }
+            | CellId::Create(_)
+            | CellId::Coordinate => None,
         }
     }
 
diff --git a/crates/improvise-wasm-client/src/render.rs b/crates/improvise-wasm-client/src/render.rs
index fa48a6bd..8aa220b4 100644
--- a/crates/improvise-wasm-client/src/render.rs
+++ b/crates/improvise-wasm-client/src/render.rs
@@ -144,7 +144,16 @@ impl RenderCell {
             text: cell.text.clone(),
             numeric: cell.numeric,
             error: cell.error,
-            editable: matches!(cell.id, Some(CellId::Record(_))),
+            editable: match cell.id {
+                Some(CellId::Record(_)) => true,
+                Some(
+                    CellId::Aggregate(_)
+                    | CellId::Derived { .. }
+                    | CellId::Create(_)
+                    | CellId::Coordinate,
+                )
+                | None => false,
+            },
             selected,
             edit_text: cell.edit_text.clone().unwrap_or_default(),
         }
@@ -563,9 +572,8 @@ mod tests {
 
     /// A 2x3 grid whose cells cover the four id shapes plus a padded slot, so
     /// one fixture answers every editability question.
-    fn group(cat: &str, name: &str, collapsed: bool) -> FrameEntry {
+    fn group(name: &str, collapsed: bool) -> FrameEntry {
         FrameEntry::Group(improvise_protocol::FrameGroup {
-            cat: cat.into(),
             name: name.into(),
             collapsed,
         })
@@ -583,11 +591,11 @@ mod tests {
             first: 0,
             cats: vec!["Month".into()],
             entries: vec![
-                group("Month", "Q1", false),
+                group("Q1", false),
                 item("Jan"),
-                group("Fiscal", "Q1", true),
+                group("Q1", true),
                 item("Feb"),
-                group("Month", "Q2", false),
+                group("Q2", false),
                 item("Apr"),
             ],
         };
@@ -628,6 +636,10 @@ mod tests {
                         }),
                     ),
                     ((1, 0), CellDisplay::text("pending")),
+                    (
+                        (1, 2),
+                        CellDisplay::text("North").with_id(CellId::Coordinate),
+                    ),
                 ])),
             ],
         });
@@ -659,6 +671,10 @@ mod tests {
             !model.rows[1].cells()[1].editable,
             "and neither is a slot fit_to_labels padded"
         );
+        assert!(
+            !model.rows[1].cells()[2].editable,
+            "renaming a coordinate moves the whole row, which is not a cell write"
+        );
     }
 
     /// The numeric/error flags are the server's answers, echoed — so the test
diff --git a/harness/tests/shell.spec.js b/harness/tests/shell.spec.js
index 9e393226..73feb294 100644
--- a/harness/tests/shell.spec.js
+++ b/harness/tests/shell.spec.js
@@ -187,6 +187,21 @@ test.describe("the standalone deployment", () => {
     await expect(page.locator("#grid td.editable").first()).toBeVisible();
   });
 
+  // improvise-i0v9: "in-page worker · <doc>" was written at 13ms, before the
+  // worker had answered anything, and read as success over a blank page.
+  test("the worker is not called up until its first frame has arrived", async ({ page }) => {
+    const doc = `first-frame-${Date.now()}.improv`;
+    await page.addInitScript(logStatus);
+    await page.goto(`/app/index.html?doc=${doc}`);
+    await expect(page.locator("#grid table")).toBeVisible({ timeout: 20_000 });
+    await expect(status(page)).toHaveText(`in-page worker · ${doc}`);
+    const log = await page.evaluate(() => window.__statusLog);
+    expect(log, `#status said: ${json(log)}`).toEqual([
+      `starting the worker · ${doc}…`,
+      `in-page worker · ${doc}`,
+    ]);
+  });
+
   // improvise-9cma: a `?doc=` this backend cannot hold used to reach the
   // generic watchdog 8s later. The REASON now crosses the port (improvise-1wft).
   test("an invalid ?doc= is named on screen, and not blamed on the browser", async ({
@@ -235,6 +250,7 @@ test.describe("the standalone deployment", () => {
 
     const second = await context.newPage();
     await second.clock.install();
+    await second.addInitScript(logStatus);
     await second.goto(`/app/index.html?doc=${doc}`);
 
     // It PAINTS. That is the whole point: a locked document is read-only, not
@@ -242,6 +258,15 @@ test.describe("the standalone deployment", () => {
     await expect(second.locator("#grid table")).toBeVisible({ timeout: 20_000 });
     await expect(second.locator("#status")).toContainText("read-only");
     await expect(second.locator("#status")).toContainText("another tab");
+    // The page's clock has not been walked, so a reason that arrived AFTER the
+    // watchdog would have been logged behind "did not answer". Nor was success
+    // ever claimed: the Snapshot that follows the notice must not overwrite it.
+    const early = await second.evaluate(() => window.__statusLog);
+    expect(early.at(-1), `#status said: ${json(early)}`).toContain("read-only");
+    expect(
+      early.filter((t) => t.includes("did not answer") || t === `in-page worker · ${doc}`),
+      `#status said: ${json(early)}`,
+    ).toEqual([]);
 
     // And the reason STANDS: it is not replaced 8s later by a watchdog
     // blaming the browser for something the code knew precisely.
@@ -252,6 +277,10 @@ test.describe("the standalone deployment", () => {
     await second.clock.runFor(HANDSHAKE_TIMEOUT_MS + 2_000);
     await expect(second.locator("#status")).toContainText("read-only");
     await expect(second.locator("#status")).not.toContainText("did not answer");
+    const late = await second.evaluate(() => window.__statusLog);
+    expect(late.some((t) => t.includes("did not answer")), `#status said: ${json(late)}`).toBe(
+      false,
+    );
 
     await second.close();
     await first.close();
diff --git a/web/app.js b/web/app.js
index 22c487ce..22e94a34 100644
--- a/web/app.js
+++ b/web/app.js
@@ -181,7 +181,8 @@ function workerTransport() {
     shell.fail(`worker failed to start: ${event.message ?? "see console"}`);
     console.error("improvise: worker error", event);
   });
-  shell.setShell(`in-page worker · ${DOC}`);
+  // Not yet a claim that it works: the first frame makes that (improvise-i0v9).
+  shell.setShell(`starting the worker · ${DOC}…`);
   // Frames posted before the worker has finished `start()` are held by
   // worker.js in arrival order, so the handshake can go out immediately.
   return { send: (text) => worker.postMessage(text), ready: Promise.resolve() };
@@ -801,6 +802,7 @@ shell.arm(
   WS_URL
     ? `${WS_URL} never answered — is improvise-ws-server running, and did it send a document?`
     : "the in-page worker did not answer (see console)",
+  { answer: WS_URL ? null : `in-page worker · ${DOC}` },
 );
 await transport.ready;
 sendViewport(); // the handshake: the snapshot is built at this extent
diff --git a/web/shell.js b/web/shell.js
index c2393afc..169373a7 100644
--- a/web/shell.js
+++ b/web/shell.js
@@ -339,7 +339,12 @@ export function createShell({
   clearTimer = clearTimeout,
 }) {
   let watchdog = null;
+  let answered = null;
   let arms = 0;
+  const disarm = () => {
+    clearTimer(watchdog);
+    watchdog = null;
+  };
 
   return {
     setShell(text) {
@@ -350,24 +355,25 @@ export function createShell({
     },
     /** A peer that never answers the handshake. Deliberately about the *first
      *  frame*, not either peer's own failure modes — clients.md. */
-    arm(message) {
+    arm(message, { answer = null } = {}) {
       arms += 1;
-      clearTimer(watchdog);
+      answered = answer;
+      disarm();
       watchdog = setTimer(() => {
         watchdog = null;
         this.setShell(message);
       }, timeoutMs);
     },
-    /** A frame arrived, so the peer answered: the watchdog has no question
-     *  left to ask. */
+    /** A frame arrived, so the peer answered. Only the FIRST, and only if no
+     *  failure spoke first, may say so (improvise-i0v9, clients.md). */
     sawFrame() {
-      clearTimer(watchdog);
-      watchdog = null;
+      if (watchdog !== null && answered !== null) this.setShell(answered);
+      disarm();
     },
     /** A peer that failed *loudly* has already said something more precise, so
      *  the watchdog must be DISARMED and not merely pre-empted (PR #13). */
     fail(text) {
-      this.sawFrame();
+      disarm();
       this.setShell(text);
     },
     armed: () => watchdog !== null,
diff --git a/web/tests/shell.test.js b/web/tests/shell.test.js
index b2d02999..e8a9570d 100644
--- a/web/tests/shell.test.js
+++ b/web/tests/shell.test.js
@@ -351,6 +351,43 @@ test("a notice DISARMS the watchdog rather than being replaced by it", () => {
   );
 });
 
+// improvise-i0v9: "in-page worker · <doc>" was written before the worker had
+// answered anything, and read as success over a blank page for eight seconds.
+test("only the first frame claims the peer answered, and not before it", () => {
+  const clock = fakeClock();
+  const seen = [];
+  const shell = createShell({
+    onShell: (t) => seen.push(t),
+    timeoutMs: 8000,
+    setTimer: clock.setTimer,
+    clearTimer: clock.clearTimer,
+  });
+  shell.arm("did not answer", { answer: "in-page worker · a.improv" });
+  assert.deepEqual(seen, [], "the claim was made before any frame arrived");
+  shell.sawFrame();
+  shell.sawFrame();
+  assert.deepEqual(seen, ["in-page worker · a.improv"]);
+});
+
+// The order worker.js really produces on a locked document: the ReadOnly
+// notice is posted after `start()` and BEFORE the Snapshot answers the handshake.
+test("a reason given before the first frame is not replaced by the frame's claim", () => {
+  const clock = fakeClock();
+  const seen = [];
+  const shell = createShell({
+    onShell: (t) => seen.push(t),
+    timeoutMs: 8000,
+    setTimer: clock.setTimer,
+    clearTimer: clock.clearTimer,
+  });
+  shell.arm("did not answer", { answer: "in-page worker · a.improv" });
+  shell.fail(noticeText({ kind: "ReadOnly", document: "a.improv", detail: "" }));
+  shell.sawFrame();
+  clock.fire();
+  assert.equal(seen.length, 1, `#status said: ${JSON.stringify(seen)}`);
+  assert.match(seen[0], /read-only/);
+});
+
 // The notice's remembered state. The value crosses `localStorage`, which is
 // allowed to forget and allowed to throw, so the interesting inputs are the
 // ones that are NOT either spelling.